Rev 3: Use pb.note() and add a bit more logging. in http://bzr.arbash-meinel.com/plugins/test_graph

[John Arbash Meinel]

At http://bzr.arbash-meinel.com/plugins/test_graph

------------------------------------------------------------
revno: 3
revision-id:john at arbash-meinel.com-20071211013859-953i0nvxzf5fo5iy
parent: john at arbash-meinel.com-20071210210105-afwkl4a38ikmbtzr
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: test_graph
timestamp: Mon 2007-12-10 19:38:59 -0600
message:
  Use pb.note() and add a bit more logging.
modified:
  graph_testing.py               graph_testing.py-20071210194758-1pwa1q7e3wnjf418-2
-------------- next part --------------
=== modified file 'graph_testing.py'
--- a/graph_testing.py	2007-12-10 21:01:05 +0000
+++ b/graph_testing.py	2007-12-11 01:38:59 +0000
@@ -22,6 +22,7 @@
     branch,
     errors,
     tsort,
+    ui,
     )
 
 
@@ -45,11 +46,16 @@
         comes first.)
         """
         interesting = []
-        for seq_num, revision_id, depth, eom in tsort.merge_sort(self.rev_graph,
-                                                                 last_revision):
-            parent_ids = self.rev_graph[revision_id]
-            if len(parent_ids) > 1:
-                interesting.append(revision_id)
+        pb = ui.ui_factory.nested_progress_bar()
+        try:
+            for seq_num, revision_id, depth, eom in tsort.merge_sort(self.rev_graph,
+                                                                     last_revision):
+                pb.update('finding interesting', seq_num, len(self.rev_graph))
+                parent_ids = self.rev_graph[revision_id]
+                if len(parent_ids) > 1:
+                    interesting.append(revision_id)
+        finally:
+            pb.finished()
         self.interesting = interesting
 
     def get_ancestry_for_interesting(self):
@@ -57,9 +63,15 @@
         ancestries = {}
         # We go in reverse order, so that we will tend to find the ancestries in
         # the map
-        for interesting_id in reversed(self.interesting):
+        pb = ui.ui_factory.nested_progress_bar()
+        cache_hit = 0
+        cache_miss = 0
+        for idx, interesting_id in enumerate(reversed(self.interesting)):
+            pb.update('finding ancestries', idx, len(self.interesting))
             needed_ids = list(self.rev_graph[interesting_id]) + [interesting_id]
             for revision_id in needed_ids:
+                if revision_id in ancestries:
+                    continue
                 ancestry = set([revision_id])
                 needed = set([revision_id])
                 while needed:
@@ -68,17 +80,23 @@
                     ancestry.update(parent_ids)
                     for parent_id in parent_ids:
                         if parent_id in ancestries:
+                            cache_hit += 1
                             ancestry.update(ancestries[parent_id])
                         else: # We need to probe in this direction
+                            cache_miss += 1
                             needed.add(parent_id)
                 ancestries[revision_id] = ancestry
+        pb.finished()
+        pb.note('Cache hit: %d, miss: %d', cache_hit, cache_miss)
         self.ancestries = ancestries
 
     def time_functions(self, a_branch):
         if self.hold_graph:
             graph = a_branch.repository.get_graph()
 
-        for revision_id in reversed(self.interesting):
+        pb = ui.ui_factory.nested_progress_bar()
+        for idx, revision_id in enumerate(reversed(self.interesting)):
+            pb.update('processing', idx, len(self.interesting))
             parent_ids = self.rev_graph[revision_id]
             a_branch.lock_read()
             try:
@@ -90,12 +108,10 @@
             finally:
                 a_branch.unlock()
             self.head_times[revision_id] = t_heads
-            parent_id_set = set(parent_ids)
-            if heads != parent_id_set:
-                if heads.issubset(parent_ids):
-                    print 'heads() is subset for: %s' % (revision_id,)
-                else:
-                    print 'heads() mismatch for: %s' % (revision_id,)
+            if heads != set(parent_ids) and heads != set(parent_ids[1:]):
+                # We frequently have nodes that are merged back into their
+                # base, but the rest should be rare
+                pb.note('heads() mismatch for: %s', revision_id)
                 self.invalid_heads.add(revision_id)
 
             a_branch.lock_read()
@@ -119,7 +135,7 @@
                 for parent_id in parent_ids:
                     left, right = graph.find_difference(revision_id, parent_id)
                     if right != set():
-                        print 'found unmerged nodes for:\n    %s\nand %s' % (
+                        pb.note('found unmerged nodes for:\n    %s\nand %s',
                             revision_id, parent_id)
                 t_trivial_diff = time.clock() - t_start
             finally:
@@ -145,13 +161,14 @@
             left_correct = left == (left_ancestry - right_ancestry)
             right_correct = right == (right_ancestry - left_ancestry)
             if not left_correct and not right_correct:
-                print 'left and right incorrect for: %s' % (revision_id,)
+                pb.note('left and right incorrect for: %s', revision_id)
             elif not left_correct:
-                print 'left incorrect for: %s' % (revision_id,)
+                pb.note('left incorrect for: %s', revision_id)
             elif not right_correct:
-                print 'right incorrect for: %s' % (revision_id,)
+                pb.note('right incorrect for: %s', revision_id)
             if not left_correct or not right_correct:
                 self.invalid_diffs.add(revision_id)
+        pb.finished()
 
     def get_statistics(self, times):
         min_time = None
@@ -184,7 +201,9 @@
     the_branch.lock_read()
     try:
         last_revision = the_branch.last_revision()
+        print 'getting revision_graph'
         rev_graph = the_branch.repository.get_revision_graph(last_revision)
+        print 'getting rev_id map'
         rev_id_map = the_branch.get_revision_id_to_revno_map()
     finally:
         the_branch.unlock()